garrytan e7c11039ea eval policy v3: judge panels gate on the per-dimension median of 3 samples ...

A judge dimension now passes when at least 2 of its 3 samples meet the
unchanged threshold (judgePanelMedian); any other sample count or a
non-numeric sample fails closed. Thresholds, sample count, judge model and
prompts are unchanged; panel logs show every sample and the mean, and name
the median as the gate. Applies to the skill-llm-eval panels, the
workflow-judge cache, the auq-matrix substance panel and the calibration
replicas. EVAL_POLICY 2 -> 3 and HARNESS_VERSION 1 -> 2 (verdicts change),
so every case starts a new pass-rate series. Replays the captured panels
from jobs 112032744416 and 112049856586 as replayed evidence.

2026-10-06 17:43:29 +00:00

..

judge-controls

C1: record workflow-judge control results (negatives separated; qa and review miss a removed step)

2026-10-04 17:58:31 +00:00

census-red.md

eval policy v3: judge panels gate on the per-dimension median of 3 samples

2026-10-06 17:43:29 +00:00

detector-inventory.md

B2: inventory of gate and periodic phrase-only detector checks

2026-10-04 18:10:57 +00:00

headroom-diagnosis-2026-10.md

B1: batching counter reads a target declared mid-line; replay corpus

2026-10-04 20:05:36 +00:00

judge-controls-2026-10.md

C1: a free test catches a removed workflow step, which the judges missed

2026-10-04 18:52:45 +00:00

judge-low-samples-2026-10.md

C2: record the judge panels rerun on the final ship and review wording

2026-10-04 18:15:27 +00:00

security-bench-ensemble-v2.json

v1.6.4.0: cut Haiku classifier FP from 44% to 23%, gate now enforced (#1135)

2026-04-23 10:23:40 -07:00